基于云理论的电力变压器故障诊断研究

针对传统的三比值法中存在的问题,根据云模型具有的随机性与模糊性特点,提出了应用云理论对三比值法进行改进的方法。该方法首先建立油中溶解气体三个比值的云模型,再通过筛选数据对模型不断地进行修正,最终结合挖掘关联规则来判定变压器故障类型,而不再局限于用固定编码定义故障类型。实例结果表明:该方法不仅有效克服了IEC三比值法边界位置误判、编码缺失、不能诊断多故障等缺点,而且还弥补了模糊三比值法忽视随机性的不足,同时使三比值法具有了自适应监督学习能力,为变压器故障诊断的准确性提供了有效的依据。

In order to solve the problems of tradition three-ratio method,this article proposes a transformer fault diagnosis model based on cloud model and three-ratio method.Firstly,the cloud model of three ratios of dissolved gas in oil is established.Then,the cloud model is constantly modified by filtering data.Eventually,it combines mining association rules to determine the transformer fault type.The results show that this method not only overcomes the shortcomings of IEC three-ratio method,such as misdiagnosis of boundary position,missing coding,inability to diagnose multiple faults,etc,but also makes up for the shortcoming of the fuzzy three-ratio method in ignoring randomness,and makes the three ratio method have the ability of self-adaptive monitoring and learning,which provides a more effective basis for transformer fault diagnosis.